Azure Blob Storage sink

Configuration

Example configurations

{
  "sinks": {
    "my_sink_id": {
      "type": "azure_blob",
      "inputs": [
        "my-source-or-transform-id"
      ],
      "container_name": "my-logs"
    }
  }
}

[sinks.my_sink_id]
type = "azure_blob"
inputs = [ "my-source-or-transform-id" ]
container_name = "my-logs"

sinks:
  my_sink_id:
    type: azure_blob
    inputs:
      - my-source-or-transform-id
    container_name: my-logs

{
  "sinks": {
    "my_sink_id": {
      "type": "azure_blob",
      "inputs": [
        "my-source-or-transform-id"
      ],
      "blob_prefix": "blob/%F/",
      "compression": "gzip",
      "connection_string": "DefaultEndpointsProtocol=https;AccountName=mylogstorage;AccountKey=storageaccountkeybase64encoded;EndpointSuffix=core.windows.net",
      "container_name": "my-logs",
      "endpoint": "https://test.blob.core.usgovcloudapi.net/",
      "storage_account": "mylogstorage"
    }
  }
}

[sinks.my_sink_id]
type = "azure_blob"
inputs = [ "my-source-or-transform-id" ]
blob_prefix = "blob/%F/"
compression = "gzip"
connection_string = "DefaultEndpointsProtocol=https;AccountName=mylogstorage;AccountKey=storageaccountkeybase64encoded;EndpointSuffix=core.windows.net"
container_name = "my-logs"
endpoint = "https://test.blob.core.usgovcloudapi.net/"
storage_account = "mylogstorage"

sinks:
  my_sink_id:
    type: azure_blob
    inputs:
      - my-source-or-transform-id
    blob_prefix: blob/%F/
    compression: gzip
    connection_string: DefaultEndpointsProtocol=https;AccountName=mylogstorage;AccountKey=storageaccountkeybase64encoded;EndpointSuffix=core.windows.net
    container_name: my-logs
    endpoint: https://test.blob.core.usgovcloudapi.net/
    storage_account: mylogstorage

acknowledgements

optional object

Controls how acknowledgements are handled for this sink.

See End-to-end Acknowledgements for more information on how event acknowledgement is handled.

acknowledgements.enabled

optional bool

Whether or not end-to-end acknowledgements are enabled.

When enabled for a sink, any source connected to that sink, where the source supports end-to-end acknowledgements as well, waits for events to be acknowledged by all connected sinks before acknowledging them at the source.

Enabling or disabling acknowledgements at the sink level takes precedence over any global acknowledgements configuration.

batch

optional object

Event batching behavior.

batch.max_bytes

optional uint

The maximum size of a batch that is processed by a sink.

This is based on the uncompressed size of the batched events, before they are serialized/compressed.

default: 1e+07 (bytes)

batch.max_events

optional uint

The maximum size of a batch before it is flushed.

batch.timeout_secs

optional float

The maximum age of a batch before it is flushed.

default: 300 (seconds)

blob_append_uuid

optional bool

Whether or not to append a UUID v4 token to the end of the blob key.

The UUID is appended to the timestamp portion of the object key, such that if the blob key generated is date=2022-07-18/1658176486, setting this field to true results in an blob key that looks like date=2022-07-18/1658176486-30f6652c-71da-4f9f-800d-a1189c47c547.

This ensures there are no name collisions, and can be useful in high-volume workloads where blob keys must be unique.

blob_prefix

optional string template

A prefix to apply to all blob keys.

Prefixes are useful for partitioning objects, such as by creating a blob key that stores blobs under a particular directory. If using a prefix for this purpose, it must end in / to act as a directory path. A trailing / is not automatically added.

Note: This parameter supports Vector's template syntax, which enables you to use dynamic per-event values.

Examples

"date/%F/hour/%H/"

"year=%Y/month=%m/day=%d/"

"kubernetes/{{ metadata.cluster }}/{{ metadata.application_name }}/"

default: blob/%F/

blob_time_format

optional string strftime

The timestamp format for the time component of the blob key.

By default, blob keys are appended with a timestamp that reflects when the blob are sent to Azure Blob Storage, such that the resulting blob key is functionally equivalent to joining the blob prefix with the formatted timestamp, such as date=2022-07-18/1658176486.

This would represent a blob_prefix set to date=%F/ and the timestamp of Mon Jul 18 2022 20:34:44 GMT+0000, with the filename_time_format being set to %s, which renders timestamps in seconds since the Unix epoch.

Supports the common strftime specifiers found in most languages.

When set to an empty string, no timestamp is appended to the blob prefix.

buffer

optional object

Configures the buffering behavior for this sink.

More information about the individual buffer types, and buffer behavior, can be found in the Buffering Model section.

buffer.max_events

optional uint

The maximum number of events allowed in the buffer.

Relevant when: type = "memory"

default: 500

buffer.max_size

required uint

The maximum size of the buffer on disk.

Must be at least ~256 megabytes (268435488 bytes).

Relevant when: type = "disk"

buffer.type

optional string literal enum

The type of buffer to use.

Enum options

Option Description

Option	Description
`disk`	Events are buffered on disk. This is less performant, but more durable. Data that has been synchronized to disk will not be lost if Vector is restarted forcefully or crashes. Data is synchronized to disk every 500ms.
`memory`	Events are buffered in memory. This is more performant, but less durable. Data will be lost if Vector is restarted forcefully or crashes.

disk

Events are buffered on disk.

This is less performant, but more durable. Data that has been synchronized to disk will not be lost if Vector is restarted forcefully or crashes.

Data is synchronized to disk every 500ms.

memory

Events are buffered in memory.

This is more performant, but less durable. Data will be lost if Vector is restarted forcefully or crashes.

default: memory

buffer.when_full

optional string literal enum

Event handling behavior when a buffer is full.

Enum options

Option Description

Option	Description
`block`	Wait for free space in the buffer. This applies backpressure up the topology, signalling that sources should slow down the acceptance/consumption of events. This means that while no data is lost, data will pile up at the edge.
`drop_newest`	Drops the event instead of waiting for free space in buffer. The event will be intentionally dropped. This mode is typically used when performance is the highest priority, and it is preferable to temporarily lose events rather than cause a slowdown in the acceptance/consumption of events.

block

Wait for free space in the buffer.

This applies backpressure up the topology, signalling that sources should slow down the acceptance/consumption of events. This means that while no data is lost, data will pile up at the edge.

drop_newest

Drops the event instead of waiting for free space in buffer.

The event will be intentionally dropped. This mode is typically used when performance is the highest priority, and it is preferable to temporarily lose events rather than cause a slowdown in the acceptance/consumption of events.

default: block

compression

optional string literal enum

Compression configuration.

All compression algorithms use the default compression level unless otherwise specified.

Enum options string literal

Option	Description
`gzip`	Gzip compression.
`none`	No compression.
`snappy`	Snappy compression.
`zlib`	Zlib compression.
`zstd`	Zstandard compression.

default: gzip

connection_string

optional string literal

The Azure Blob Storage Account connection string.

Authentication with access key is the only supported authentication method.

Either storage_account, or this field, must be specified.

Examples

"DefaultEndpointsProtocol=https;AccountName=mylogstorage;AccountKey=storageaccountkeybase64encoded;EndpointSuffix=core.windows.net"

container_name

required string literal

The Azure Blob Storage Account container name.

Examples

"my-logs"

encoding

required object

Configures how events are encoded into raw bytes.

encoding.avro

required object

Apache Avro-specific encoder options.

Relevant when: codec = "avro"

encoding.avro.schema

required string literal

The Avro schema.

Examples

"{ \"type\": \"record\", \"name\": \"log\", \"fields\": [{ \"name\": \"message\", \"type\": \"string\" }] }"

encoding.codec

required string literal enum

The codec to use for encoding events.

Enum options

Option	Description
`avro`	Encodes an event as an Apache Avro message.
`csv`	Encodes an event as a CSV message. This codec must be configured with fields to encode.
`gelf`	Encodes an event as a GELF message. This codec is experimental for the following reason: The GELF specification is more strict than the actual Graylog receiver. Vector’s encoder currently adheres more strictly to the GELF spec, with the exception that some characters such as `@` are allowed in field names. Other GELF codecs such as Loki’s, use a Go SDK that is maintained by Graylog, and is much more relaxed than the GELF spec. Going forward, Vector will use that Go SDK as the reference implementation, which means the codec may continue to relax the enforcement of specification.
`json`	Encodes an event as JSON.
`logfmt`	Encodes an event as a logfmt message.
`native`	Encodes an event in the native Protocol Buffers format. This codec is experimental.
`native_json`	Encodes an event in the native JSON format. This codec is experimental.
`protobuf`	Encodes an event as a Protobuf message.
`raw_message`	No encoding. This encoding uses the `message` field of a log event. Be careful if you are modifying your log events (for example, by using a `remap` transform) and removing the message field while doing additional parsing on it, as this could lead to the encoding emitting empty strings for the given event.
`text`	Plain text encoding. This encoding uses the `message` field of a log event. For metrics, it uses an encoding that resembles the Prometheus export format. Be careful if you are modifying your log events (for example, by using a `remap` transform) and removing the message field while doing additional parsing on it, as this could lead to the encoding emitting empty strings for the given event.

Examples

"avro"

"csv"

"gelf"

"json"

"logfmt"

"native"

"native_json"

"protobuf"

"raw_message"

"text"

encoding.csv

required object

The CSV Serializer Options.

Relevant when: codec = "csv"

encoding.csv.capacity

optional uint

Set the capacity (in bytes) of the internal buffer used in the CSV writer. This defaults to a reasonable setting.

default: 8192

encoding.csv.delimiter

optional ascii_char

The field delimiter to use when writing CSV.

default: ,

encoding.csv.double_quote

optional bool

Enable double quote escapes.

This is enabled by default, but it may be disabled. When disabled, quotes in field data are escaped instead of doubled.

default: true

encoding.csv.escape

optional ascii_char

The escape character to use when writing CSV.

In some variants of CSV, quotes are escaped using a special escape character like \ (instead of escaping quotes by doubling them).

To use this, double_quotes needs to be disabled as well otherwise it is ignored.

default: "

encoding.csv.fields

required [string]

Configures the fields that will be encoded, as well as the order in which they appear in the output.

If a field is not present in the event, the output will be an empty string.

Values of type Array, Object, and Regex are not supported and the output will be an empty string.

encoding.csv.quote

optional ascii_char

The quote character to use when writing CSV.

default: "

encoding.csv.quote_style

optional string literal enum

The quoting style to use when writing CSV data.

Enum options

Option	Description
`always`	Always puts quotes around every field.
`necessary`	Puts quotes around fields only when necessary. They are necessary when fields contain a quote, delimiter, or record terminator. Quotes are also necessary when writing an empty record (which is indistinguishable from a record with one empty field).
`never`	Never writes quotes, even if it produces invalid CSV data.
`non_numeric`	Puts quotes around all fields that are non-numeric. Namely, when writing a field that does not parse as a valid float or integer, then quotes are used even if they aren’t strictly necessary.

default: necessary

encoding.except_fields

optional [string]

List of fields that are excluded from the encoded event.

encoding.json

optional object

Options for the JsonSerializer.

Relevant when: codec = "json"

encoding.json.pretty

optional bool

Whether to use pretty JSON formatting.

default: false

encoding.metric_tag_values

optional string literal enum

Controls how metric tag values are encoded.

When set to single, only the last non-bare value of tags are displayed with the metric. When set to full, all metric tags are exposed as separate assignments.

Relevant when: codec = "json" or codec = "text"

Enum options

Option	Description
`full`	All tags are exposed as arrays of either string or null values.
`single`	Tag values are exposed as single strings, the same as they were before this config option. Tags with multiple values show the last assigned value, and null values are ignored.

default: single

encoding.only_fields

optional [string]

List of fields that are included in the encoded event.

encoding.protobuf

required object

Options for the Protobuf serializer.

Relevant when: codec = "protobuf"

encoding.protobuf.desc_file

required string literal

The path to the protobuf descriptor set file.

This file is the output of protoc -o <path> ...

Examples

"/etc/vector/protobuf_descriptor_set.desc"

encoding.protobuf.message_type

required string literal

The name of the message type to use for serializing.

Examples

"package.Message"

encoding.timestamp_format

optional string literal enum

Format used for timestamp fields.

Enum options

Option	Description
`rfc3339`	Represent the timestamp as a RFC 3339 timestamp.
`unix`	Represent the timestamp as a Unix timestamp.
`unix_float`	Represent the timestamp as a Unix timestamp in floating point.
`unix_ms`	Represent the timestamp as a Unix timestamp in milliseconds.
`unix_ns`	Represent the timestamp as a Unix timestamp in nanoseconds.
`unix_us`	Represent the timestamp as a Unix timestamp in microseconds

endpoint

optional string literal

The Azure Blob Storage Endpoint URL.

This is used to override the default blob storage endpoint URL in cases where you are using credentials read from the environment/managed identities or access tokens without using an explicit connection_string (which already explicitly supports overriding the blob endpoint URL).

This may only be used with storage_account and is ignored when used with connection_string.

Examples

"https://test.blob.core.usgovcloudapi.net/"

"https://test.blob.core.windows.net/"

framing

optional object

Framing configuration.

framing.character_delimited

required object

Options for the character delimited encoder.

Relevant when: method = "character_delimited"

framing.character_delimited.delimiter

required ascii_char

The ASCII (7-bit) character that delimits byte sequences.

framing.length_delimited

required object

Options for the length delimited decoder.

Relevant when: method = "length_delimited"

framing.length_delimited.length_field_is_big_endian

optional bool

Length field byte order (little or big endian)

default: true

framing.length_delimited.length_field_length

optional uint

Number of bytes representing the field length

default: 4

framing.length_delimited.length_field_offset

optional uint

Number of bytes in the header before the length field

framing.length_delimited.max_frame_length

optional uint

Maximum frame length

default: 8.388608e+06

framing.method

required string literal enum

The framing method.

Enum options

Option	Description
`bytes`	Event data is not delimited at all.
`character_delimited`	Event data is delimited by a single ASCII (7-bit) character.
`length_delimited`	Event data is prefixed with its length in bytes. The prefix is a 32-bit unsigned integer, little endian.
`newline_delimited`	Event data is delimited by a newline (LF) character.

Examples

"bytes"

"character_delimited"

"length_delimited"

"newline_delimited"

healthcheck

optional object

Healthcheck configuration.

healthcheck.enabled

optional bool

Whether or not to check the health of the sink when Vector starts up.

default: true

inputs

required [string]

A list of upstream source or transform IDs.

Wildcards (*) are supported.

See configuration for more info.

Array string literal

Examples

[
  "my-source-or-transform-id",
  "prefix-*"
]

request

optional object

Middleware settings for outbound requests.

Various settings can be configured, such as concurrency and rate limits, timeouts, retry behavior, etc.

Note that the retry backoff policy follows the Fibonacci sequence.

request.adaptive_concurrency

optional object

Configuration of adaptive concurrency parameters.

These parameters typically do not require changes from the default, and incorrect values can lead to meta-stable or unstable performance and sink behavior. Proceed with caution.

request.adaptive_concurrency.decrease_ratio

optional float

The fraction of the current value to set the new concurrency limit when decreasing the limit.

Valid values are greater than 0 and less than 1. Smaller values cause the algorithm to scale back rapidly when latency increases.

Note that the new limit is rounded down after applying this ratio.

default: 0.9

request.adaptive_concurrency.ewma_alpha

optional float

The weighting of new measurements compared to older measurements.

Valid values are greater than 0 and less than 1.

ARC uses an exponentially weighted moving average (EWMA) of past RTT measurements as a reference to compare with the current RTT. Smaller values cause this reference to adjust more slowly, which may be useful if a service has unusually high response variability.

default: 0.4

request.adaptive_concurrency.initial_concurrency

optional uint

The initial concurrency limit to use. If not specified, the initial limit will be 1 (no concurrency).

It is recommended to set this value to your service’s average limit if you’re seeing that it takes a long time to ramp up adaptive concurrency after a restart. You can find this value by looking at the adaptive_concurrency_limit metric.

default: 1

request.adaptive_concurrency.max_concurrency_limit

optional uint

The maximum concurrency limit.

The adaptive request concurrency limit will not go above this bound. This is put in place as a safeguard.

default: 200

request.adaptive_concurrency.rtt_deviation_scale

optional float

Scale of RTT deviations which are not considered anomalous.

Valid values are greater than or equal to 0, and we expect reasonable values to range from 1.0 to 3.0.

When calculating the past RTT average, we also compute a secondary “deviation” value that indicates how variable those values are. We use that deviation when comparing the past RTT average to the current measurements, so we can ignore increases in RTT that are within an expected range. This factor is used to scale up the deviation to an appropriate range. Larger values cause the algorithm to ignore larger increases in the RTT.

default: 2.5

request.concurrency

optional string literal enum uint

Configuration for outbound request concurrency.

This can be set either to one of the below enum values or to a positive integer, which denotes a fixed concurrency limit.

Enum options

Option Description

adaptive Concurrency will be managed by Vector’s Adaptive Request Concurrency feature.

Option	Description
`adaptive`	Concurrency will be managed by Vector’s Adaptive Request Concurrency feature.
`none`	A fixed concurrency of 1. Only one request can be outstanding at any given time.

none

A fixed concurrency of 1.

Only one request can be outstanding at any given time.

default: adaptive

request.rate_limit_duration_secs

optional uint

The time window used for the rate_limit_num option.

default: 1 (seconds)

request.rate_limit_num

optional uint

The maximum number of requests allowed within the rate_limit_duration_secs time window.

default: 250 (requests)

request.retry_attempts

optional uint

The maximum number of retries to make for failed requests.

default: 9.223372036854776e+18 (retries)

request.retry_initial_backoff_secs

optional uint

The amount of time to wait before attempting the first retry for a failed request.

After the first retry has failed, the fibonacci sequence is used to select future backoffs.

default: 1 (seconds)

request.retry_jitter_mode

optional string literal enum

The jitter mode to use for retry backoff behavior.

Enum options

Option Description

Option	Description
`Full`	Full jitter. The random delay is anywhere from 0 up to the maximum current delay calculated by the backoff strategy. Incorporating full jitter into your backoff strategy can greatly reduce the likelihood of creating accidental denial of service (DoS) conditions against your own systems when many clients are recovering from a failure state.
`None`	No jitter.

Full

Full jitter.

The random delay is anywhere from 0 up to the maximum current delay calculated by the backoff strategy.

Incorporating full jitter into your backoff strategy can greatly reduce the likelihood of creating accidental denial of service (DoS) conditions against your own systems when many clients are recovering from a failure state.

None No jitter.

default: Full

request.retry_max_duration_secs

optional uint

The maximum amount of time to wait between retries.

default: 30 (seconds)

request.timeout_secs

optional uint

The time a request can take before being aborted.

Datadog highly recommends that you do not lower this value below the service’s internal timeout, as this could create orphaned requests, pile on retries, and result in duplicate data downstream.

default: 60 (seconds)

storage_account

optional string literal

The Azure Blob Storage Account name.

Attempts to load credentials for the account in the following ways, in order:

read from environment variables (more information)
looks for a Managed Identity
uses the az CLI tool to get an access token (more information)

Either connection_string, or this field, must be specified.

Examples

"mylogstorage"

Telemetry

intentional

True if the events were discarded intentionally, like a filter transform, or false if due to an error.

pid optional

The process ID of the Vector instance.

component_errors_total

counter

The total number of errors encountered by this component.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

error_type

The type of the error

host optional

The hostname of the system Vector is running on.

pid optional

The process ID of the Vector instance.

stage

The stage within the component at which the error occurred.

component_received_event_bytes_total

counter

The number of event bytes accepted by this component either from tagged origins like file and uri, or cumulatively from other origins.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

container_name optional

The name of the container from which the data originated.

file optional

The file from which the data originated.

host optional

The hostname of the system Vector is running on.

mode optional

The connection mode used by the component.

peer_addr optional

The IP from which the data originated.

peer_path optional

The pathname from which the data originated.

pid optional

The process ID of the Vector instance.

pod_name optional

The name of the pod from which the data originated.

uri optional

The sanitized URI from which the data originated.

component_received_events_count

histogram

A histogram of the number of events passed in each internal batch in Vector’s internal topology.

Note that this is separate than sink-level batching. It is mostly useful for low level debugging performance issues in Vector due to small internal batches.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

container_name optional

The name of the container from which the data originated.

file optional

The file from which the data originated.

host optional

The hostname of the system Vector is running on.

mode optional

The connection mode used by the component.

peer_addr optional

The IP from which the data originated.

peer_path optional

The pathname from which the data originated.

pid optional

The process ID of the Vector instance.

pod_name optional

The name of the pod from which the data originated.

uri optional

The sanitized URI from which the data originated.

component_received_events_total

counter

The number of events accepted by this component either from tagged origins like file and uri, or cumulatively from other origins.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

container_name optional

The name of the container from which the data originated.

file optional

The file from which the data originated.

host optional

The hostname of the system Vector is running on.

mode optional

The connection mode used by the component.

peer_addr optional

The IP from which the data originated.

peer_path optional

The pathname from which the data originated.

pid optional

The process ID of the Vector instance.

pod_name optional

The name of the pod from which the data originated.

uri optional

The sanitized URI from which the data originated.

component_sent_bytes_total

counter

The number of raw bytes sent by this component to destination sinks.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

endpoint optional

The endpoint to which the bytes were sent. For HTTP, this will be the host and path only, excluding the query string.

file optional

The absolute path of the destination file.

host optional

The hostname of the system Vector is running on.

pid optional

The process ID of the Vector instance.

protocol

The protocol used to send the bytes.

region optional

The AWS region name to which the bytes were sent. In some configurations, this may be a literal hostname.

component_sent_event_bytes_total

counter

The total number of event bytes emitted by this component.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

host optional

The hostname of the system Vector is running on.

output optional

The specific output of the component.

pid optional

The process ID of the Vector instance.

component_sent_events_total

counter

The total number of events emitted by this component.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

host optional

The hostname of the system Vector is running on.

output optional

The specific output of the component.

pid optional

The process ID of the Vector instance.

utilization

gauge

A ratio from 0 to 1 of the load on a component. A value of 0 would indicate a completely idle component that is simply waiting for input. A value of 1 would indicate a that is never idle. This value is updated every 5 seconds.

component_id

The Vector component ID.

component_kind

The Vector component kind.

component_type

The Vector component type.

host optional

The hostname of the system Vector is running on.

pid optional

The process ID of the Vector instance.

How it works

Buffers and batches

This component buffers & batches data as shown in the diagram above. You’ll notice that Vector treats these concepts differently, instead of treating them as global concepts, Vector treats them as sink specific concepts. This isolates sinks, ensuring services disruptions are contained and delivery guarantees are honored.

Batches are flushed when 1 of 2 conditions are met:

The batch age meets or exceeds the configured timeout_secs.
The batch size meets or exceeds the configured max_bytes or max_events.

Buffers are controlled via the buffer.* options.

Health checks

Health checks ensure that the downstream service is accessible and ready to accept data. This check is performed upon sink initialization. If the health check fails an error will be logged and Vector will proceed to start.

Require health checks

If you’d like to exit immediately upon a health check failure, you can pass the --require-healthy flag:

vector --config /etc/vector/vector.yaml --require-healthy

Disable health checks

If you’d like to disable health checks for this sink you can set the healthcheck option to false.

Object naming

By default, Vector names your blobs different based on whether or not the blobs are compressed.

Here is the format without compression:

<key_prefix><timestamp>-<uuidv4>.log

Here’s an example blob name without compression:

blob/2021-06-23/1560886634-fddd7a0e-fad9-4f7e-9bce-00ae5debc563.log

And here is the format with compression:

<key_prefix><timestamp>-<uuidv4>.log.gz

An example blob name with compression:

blob/2021-06-23/1560886634-fddd7a0e-fad9-4f7e-9bce-00ae5debc563.log.gz

Vector appends a UUIDV4 token to ensure there are no name conflicts in the unlikely event that two Vector instances are writing data at the same time.

You can control the resulting name via the blob_prefix, blob_time_format, and blob_append_uuid options.

For example, to store objects at the root Azure storage folder, without a timestamp or UUID use these configuration options:

blob_prefix = "{{ my_file_name }}"
blob_time_format = ""
blob_append_uuid = false

Rate limits & adaptive concurrency

Adaptive Request Concurrency (ARC)

Adaptive Request Concurrency is a feature of Vector that does away with static concurrency limits and automatically optimizes HTTP concurrency based on downstream service responses. The underlying mechanism is a feedback loop inspired by TCP congestion control algorithms. Checkout the announcement blog post,

We highly recommend enabling this feature as it improves performance and reliability of Vector and the systems it communicates with. As such, we have made it the default, and no further configuration is required.

Static concurrency

If Adaptive Request Concurrency is not for you, you can manually set static concurrency limits by specifying an integer for request.concurrency:

sinks:
	my-sink:
		request:
			concurrency: 10

Rate limits

In addition to limiting request concurrency, you can also limit the overall request throughput via the request.rate_limit_duration_secs and request.rate_limit_num options.

sinks:
	my-sink:
		request:
			rate_limit_duration_secs: 1
			rate_limit_num: 10

These will apply to both adaptive and fixed request.concurrency values.

Retry policy

Vector will retry failed requests (status == 429, >= 500, and != 501). Other responses will not be retried. You can control the number of retry attempts and backoff rate with the request.retry_attempts and request.retry_backoff_secs options.

State

This component is stateless, meaning its behavior is consistent across each input.